# import libraries

# basic
import warnings
warnings.filterwarnings('ignore')

import os
import random
import pydicom
import itertools
import numpy as np
import pandas as pd
from sklearn.preprocessing import MultiLabelBinarizer
from skmultilearn.model_selection import iterative_train_test_split

# visualization
from PIL import Image
import matplotlib.pyplot as plt
import matplotlib.patches as patches
import matplotlib.gridspec as gridspec

# object detection
import json
from skimage.measure import label as sk_label
from skimage.measure import regionprops as sk_regions
from torchvision.transforms.v2 import functional as F

class config:
    
    root = "/kaggle/input/hwk05-data/hwk05_data" ## your own base root path
    seed = 42

def seed_everything(seed):
    # Set Python random seed
    random.seed(seed)
    
    # Set NumPy random seed
    np.random.seed(seed)
    
seed_everything(config.seed)

# training dataframe
train_df = pd.read_csv("/kaggle/input/hwk05-data/hwk05_data/train/train.csv", encoding='big5')
train_df

# all classes
category = {
    "心臟肥大": "cardiac_hypertrophy",
    "主動脈硬鈣化": "aortic_atherosclerosis_calcification",
    "主動脈彎曲": "aortic_curvature",
    "肺尖肋膜增厚": "intercostal_pleural_thickening",
    "肺野浸潤增加": "lung_field_infiltration",
    "胸椎退化性關節病變": "degenerative_joint_disease_of_the_thoracic_spine",
    "脊椎側彎": "scoliosis",
    "normal": "normal"
}

# change category names to English
def change_to_eng_names(df):
    
    df["category"] = df["category"].apply(lambda x: category[x])
    df["ImagePath"] = df.apply(lambda df: "/".join([df["category"], "image", df["Filename"]]), axis=1)
    df["MarkPath"] = df.apply(lambda df: "/".join([df["category"], "mark", df["Filename"] + ".jpg"]), axis=1)

change_to_eng_names(train_df)
train_df

temp = train_df[train_df["category"].duplicated() == False]
temp

def plot_images_and_marks(df):
    
    temp = df[df["category"].duplicated() == False]
    
    rows, cols = 4, 2
    fig = plt.figure(figsize = (16, 16))
    grid = plt.GridSpec(rows, cols)
    
    for i in range(rows * cols):
        image = pydicom.dcmread(os.path.join(config.root, "train", temp.iloc[i, 5])).pixel_array
        if temp.iloc[i, 1] != "normal":
            mark = np.array(Image.open(os.path.join(config.root, "train", temp.iloc[i, 6])))
        else:
            mark = np.zeros((image.shape[0], image.shape[1]))
        
        categories = fig.add_subplot(grid[i])
        categories.set_title(f"{temp.iloc[i, 1]}\n", fontweight = 'semibold', size = 14)
        categories.set_axis_off()
        
        gs = gridspec.GridSpecFromSubplotSpec(1, 2, subplot_spec = grid[i])
        
        ax = fig.add_subplot(gs[0])
        ax.imshow(image, cmap = "gray")
        ax.set_title("Image")
        ax.axis("off")
        
        ax = fig.add_subplot(gs[1], sharey = ax)
        ax.imshow(mark, cmap = "gray")
        ax.set_title("Mark")
        ax.axis("off")
        
    fig.patch.set_facecolor('white')
    fig.suptitle("Images and marks of 8 categories\n", fontweight = 'bold', size = 16)
    fig.tight_layout()
      
plot_images_and_marks(train_df)

def X_ray_normalization(dcm_file, vmin, vmax):
    
    img = pydicom.dcmread(dcm_file)
    origin = img.pixel_array
    
    # needed values
    WW = img.WindowWidth
    WC = img.WindowCenter
    BitsStored = img.BitsStored

    # Compute min and max intensity bounds
    imin = WC - (WW / 2)
    imax = WC + (WW / 2)
    # Clip pixel values based on imin and imax
    clipped = np.clip(origin, imin, imax)
    # Perform intensity log-transformation
    log_img = -np.log((1 + clipped) / (2 ** BitsStored))
    
    # simplest color balance algorithm
    # lower_bound = np.percentile(log_img, vmin)
    # upper_bound = np.percentile(log_img, vmax)
    normalize_img = (log_img - vmin) / (vmax - vmin)
    normalize_img = np.clip(normalize_img, 0, 1)

    
    return origin, log_img, normalize_img

def plot_before_and_after(ID, df):
    
    patient_df = df[df["ID"] == ID]
    path = os.path.join(config.root, "train", patient_df.iloc[0, 5])
    origin, log_img, normalize_img = X_ray_normalization(path, vmin = 0, vmax = 2.5)
    
    plt.figure(figsize = (16, 16))
    fig, ax = plt.subplots(1, 3)
    np.vectorize(lambda ax: ax.axis('off'))(ax)
    plt.subplots_adjust(wspace = None, hspace = None)
    
    ax[0].imshow(origin, cmap = "gray")
    ax[0].set_title("Original Image", size = 8)
    ax[1].imshow(log_img, cmap = "gray")
    ax[1].set_title("After Log-transformation", size = 8)
    ax[2].imshow(normalize_img, cmap = "gray")
    ax[2].set_title("After Normalization", size = 8)
          
    fig.suptitle(f"{ID}", fontweight = 'bold', size = 10, x = 0.52, y = 0.77)
    
plot_before_and_after(ID = "TDR02_20161209_161439", df = train_df)

<Figure size 1600x1600 with 0 Axes>

def mask_to_bbox(mark_path):
    img = np.array(Image.open(mark_path))

    mask = img != 0
    sk_mask = sk_label(mask, connectivity = 2)
    regions = sk_regions(sk_mask)
    bboxes = []
    for region in regions:
        if region.area < 3000 :
            continue
        bboxes.append(region.bbox)

    ymin, xmin, ymax, xmax = bboxes[0]
    
    return xmin, ymin, xmax, ymax

def plot_bbox_and_mark(df):
    
    temp = df[df["category"].duplicated() == False]
    
    rows, cols = 4, 2
    fig = plt.figure(figsize = (16, 16))
    grid = plt.GridSpec(rows, cols)
    
    for i in range(rows * cols):
        
        path = os.path.join(config.root, "train", temp.iloc[i, 5])
        mark_path = os.path.join(config.root, "train", temp.iloc[i, 6])
        
        _, _, after = X_ray_normalization(path, vmin = 0, vmax = 2.5)
        
        if temp.iloc[i, 1] != "normal":
            mark = np.array(Image.open(mark_path))
            xmin, ymin, xmax, ymax = mask_to_bbox(mark_path)
        else:
            mark = np.zeros((after.shape[0], after.shape[1]))
            xmin, ymin, xmax, ymax = 0, 0, 0, 0
        
        bbox = patches.Rectangle((xmin, ymin), xmax - xmin, ymax - ymin, linewidth = 2, 
                                 edgecolor = "r", facecolor = 'none')
        
        categories = fig.add_subplot(grid[i])
        categories.set_title(f"{temp.iloc[i, 1]}\n", fontweight = 'semibold', size = 14)
        categories.set_axis_off()
        
        gs = gridspec.GridSpecFromSubplotSpec(1, 3, subplot_spec = grid[i])
        
        ax = fig.add_subplot(gs[0])
        ax.imshow(after, cmap = "gray")
        ax.set_title("Image")
        ax.axis("off")
        
        ax = fig.add_subplot(gs[1], sharey = ax)
        ax.imshow(after, cmap = "gray")
        ax.add_patch(bbox)
        ax.set_title("Image with bbox")
        ax.axis("off")
        
        ax = fig.add_subplot(gs[2], sharey = ax)
        ax.imshow(mark, cmap = "gray")
        ax.set_title("Mark")
        ax.axis("off")
        
    fig.patch.set_facecolor('white')
    fig.suptitle("Images with bbox and marks of 8 categories\n", fontweight = 'bold', size = 16)
    fig.tight_layout()
      
plot_bbox_and_mark(train_df)

def write_bbox(df):
    
    all_xmin, all_ymin, all_xmax, all_ymax = [], [], [], []
    
    for i in range(df.shape[0]):
        
        if df.iloc[i, 1] != "normal":
            mark_path = os.path.join(config.root, "train", df.iloc[i, 6])
            xmin, ymin, xmax, ymax = mask_to_bbox(mark_path)
        else:
            xmin, ymin, xmax, ymax = 0, 0, 0, 0
        
        all_xmin.append(xmin)
        all_ymin.append(ymin)
        all_xmax.append(xmax)
        all_ymax.append(ymax)
        
    df["xmin"] = all_xmin
    df["ymin"] = all_ymin
    df["xmax"] = all_xmax
    df["ymax"] = all_ymax
    
write_bbox(train_df)
train_df

labels = list(train_df["category"].unique())
label2class = {l: c for c, l in enumerate(labels)}
label2class

{'normal': 0,
 'aortic_curvature': 1,
 'aortic_atherosclerosis_calcification': 2,
 'cardiac_hypertrophy': 3,
 'intercostal_pleural_thickening': 4,
 'lung_field_infiltration': 5,
 'degenerative_joint_disease_of_the_thoracic_spine': 6,
 'scoliosis': 7}

# write class_id
def write_class_id(df):
    class_id = []
    for i in range(df.shape[0]):
        class_id.append(label2class[df.iloc[i, 1]])
    df["class_id"] = class_id
    
write_class_id(train_df)
train_df

train_df.nunique()['ID'], train_df.shape[0]

(348, 451)

binarizer = MultiLabelBinarizer()
disease_id = []
for ID in train_df.ID.unique():
    diseases = []
    temp = train_df[train_df["ID"] == ID]
    diseases.extend(list(temp["class_id"]))
    disease_id.append(diseases)

one_hot = binarizer.fit_transform(disease_id)
one_hot

array([[1, 0, 0, ..., 0, 0, 0],
       [1, 0, 0, ..., 0, 0, 0],
       [1, 0, 0, ..., 0, 0, 0],
       ...,
       [0, 0, 0, ..., 0, 0, 1],
       [0, 0, 0, ..., 0, 0, 1],
       [0, 0, 0, ..., 0, 0, 1]])

one_hot.shape

(348, 8)

train_ID, train_label, val_ID, val_label = iterative_train_test_split(np.expand_dims(train_df["ID"].unique(), axis = 1), one_hot, test_size = 0.2)

training = train_df[train_df["ID"].isin(train_ID.ravel())]
validation = train_df[train_df["ID"].isin(val_ID.ravel())]

categories = []
for l, c in label2class.items():
    if l == "normal":
        continue
    categories.append({"id": c, "name": l})

categories

[{'id': 1, 'name': 'aortic_curvature'},
 {'id': 2, 'name': 'aortic_atherosclerosis_calcification'},
 {'id': 3, 'name': 'cardiac_hypertrophy'},
 {'id': 4, 'name': 'intercostal_pleural_thickening'},
 {'id': 5, 'name': 'lung_field_infiltration'},
 {'id': 6, 'name': 'degenerative_joint_disease_of_the_thoracic_spine'},
 {'id': 7, 'name': 'scoliosis'}]

# change data to coco format
def coco_format(df, categories):
    coco_output = {
        "images" : [],
        "categories" : [],
        "annotations" : []
        }
 
    coco_output['categories'] = categories

    annotation_id = 0
    for image_id, img_name in enumerate(df.ID.unique()):
        image_df = df[df.ID == img_name]
        if len(image_df) == 1:
            image_dict = {
                "file_name" : list(image_df.category)[0] + "/" + list(image_df.Filename)[0].replace(".dcm", ".jpg"),
                "height" : int(image_df.Height),
                "width" : int(image_df.Width),
                "id" : image_id
                }
        else:
            unique = image_df.iloc[0, :]
            image_dict = {
                "file_name" : unique.category + "/" + unique.Filename.replace(".dcm", ".jpg"),
                "height" : int(unique.Height),
                "width" : int(unique.Width),
                "id" : image_id
                }  
        coco_output['images'].append(image_dict)

        for _, row in image_df.iterrows():
            xmin = int(row.xmin)
            ymin = int(row.ymin)
            xmax = int(row.xmax)
            ymax = int(row.ymax)
            if xmin == ymin == xmax == ymax == 0:
                continue

            area = (xmax - xmin) * (ymax - ymin)
          
            poly = [
                (xmin, ymin), (xmax, ymin), 
                (xmax, ymax), (xmin, ymax)
            ]
            poly = list(itertools.chain.from_iterable(poly))

            mask_dict = {
                "id" : annotation_id,
                "image_id" : image_id,
                "category_id" : row.class_id,
                "bbox" : [xmin, ymin, (xmax - xmin), (ymax - ymin)],
                "area" : area,
                "iscrowd" : 0, 
                "segmentation" : [poly],
                }
            coco_output["annotations"].append(mask_dict)
            annotation_id += 1

    return coco_output

train_coco = coco_format(training, categories)
val_coco = coco_format(validation, categories)

train_coco['images'][:5]

[{'file_name': 'normal/220_94.jpg', 'height': 2560, 'width': 2472, 'id': 0},
 {'file_name': 'normal/220_93.jpg', 'height': 2496, 'width': 2312, 'id': 1},
 {'file_name': 'normal/220_90.jpg', 'height': 2632, 'width': 2320, 'id': 2},
 {'file_name': 'normal/220_88.jpg', 'height': 2624, 'width': 2560, 'id': 3},
 {'file_name': 'normal/220_86.jpg', 'height': 2632, 'width': 2544, 'id': 4}]

train_coco['annotations'][:5]

[{'id': 0,
  'image_id': 64,
  'category_id': 1,
  'bbox': [904, 656, 760, 1143],
  'area': 868680,
  'iscrowd': 0,
  'segmentation': [[904, 656, 1664, 656, 1664, 1799, 904, 1799]]},
 {'id': 1,
  'image_id': 64,
  'category_id': 5,
  'bbox': [232, 128, 2200, 2024],
  'area': 4452800,
  'iscrowd': 0,
  'segmentation': [[232, 128, 2432, 128, 2432, 2152, 232, 2152]]},
 {'id': 2,
  'image_id': 65,
  'category_id': 1,
  'bbox': [872, 736, 632, 1072],
  'area': 677504,
  'iscrowd': 0,
  'segmentation': [[872, 736, 1504, 736, 1504, 1808, 872, 1808]]},
 {'id': 3,
  'image_id': 65,
  'category_id': 5,
  'bbox': [144, 280, 2168, 1952],
  'area': 4231936,
  'iscrowd': 0,
  'segmentation': [[144, 280, 2312, 280, 2312, 2232, 144, 2232]]},
 {'id': 4,
  'image_id': 66,
  'category_id': 1,
  'bbox': [1024, 528, 608, 1040],
  'area': 632320,
  'iscrowd': 0,
  'segmentation': [[1024, 528, 1632, 528, 1632, 1568, 1024, 1568]]}]

train_coco['categories']

[{'id': 1, 'name': 'aortic_curvature'},
 {'id': 2, 'name': 'aortic_atherosclerosis_calcification'},
 {'id': 3, 'name': 'cardiac_hypertrophy'},
 {'id': 4, 'name': 'intercostal_pleural_thickening'},
 {'id': 5, 'name': 'lung_field_infiltration'},
 {'id': 6, 'name': 'degenerative_joint_disease_of_the_thoracic_spine'},
 {'id': 7, 'name': 'scoliosis'}]

def dcm_to_jpg(df):
    
    for path in df.ImagePath:
        dcm_path = os.path.join(config.root, "train", path)
        _, _, image = X_ray_normalization(dcm_path, vmin = 0, vmax = 2.5)
        file = os.path.join("/kaggle/working/train", path.split("/")[0])
        # file = os.path.join(config.root,"kaggle/working/train", path.split("/")[0])
        jpg_name = path.split("/")[-1].replace(".dcm", ".jpg")

        if os.path.isdir(file) == False:
            os.makedirs(file)
            print("makedirs")

        plt.imsave(f"{file}/{jpg_name}", image, cmap = "gray")

dcm_to_jpg(train_df)

makedirs
makedirs
makedirs
makedirs
makedirs
makedirs
makedirs
makedirs

with open("train.json", "w") as outfile:
    json.dump(train_coco, outfile)
    

with open("val.json", "w") as outfile:
    json.dump(val_coco, outfile)

def Testdcm_to_jpg(df):
    
    for path in df.ImagePath:
        path=path.lstrip("/")
        dcm_path = os.path.join(config.root, "test", path)
        _, _, image = X_ray_normalization(dcm_path, vmin = 0, vmax = 2.5)
        file = os.path.join("/kaggle/working/test", path.split("/")[0])
        # file = os.path.join(config.root,"/kaggle/working/test", path.split("/")[0])
        jpg_name = path.split("/")[-1].replace(".dcm", ".jpg")

        if os.path.isdir(file) == False:
            os.makedirs(file)
            print("makedirs")

        plt.imsave(f"{file}/{jpg_name}", image, cmap = "gray")

test_df = pd.read_csv("/kaggle/input/hwk05-data/hwk05_data/test/test.csv", encoding='big5')
Testdcm_to_jpg(test_df)

makedirs

# import libraries

# basic
#import warnings
#warnings.filterwarnings('ignore')

import os
os.environ['CUBLAS_WORKSPACE_CONFIG'] = ':4096:8'
import random
import numpy as np
import pandas as pd
import math
from tqdm.notebook import tqdm

# visualization
from PIL import Image
import matplotlib.pyplot as plt
import matplotlib.patches as patches

# PyTorch
import torch
import torchvision
from torch.utils.data import Dataset, DataLoader
from torchvision import models
from torchvision.models.detection.faster_rcnn import FastRCNNPredictor
from torchvision.models.detection.rpn import AnchorGenerator
from torchvision.models.detection import FasterRCNN
from torchvision.transforms import v2
from torchvision import tv_tensors
from torchvision.tv_tensors import BoundingBoxes

# object detection
!pip install pycocotools
import pycocotools
from pycocotools.coco import COCO

Requirement already satisfied: pycocotools in /usr/local/lib/python3.10/dist-packages (2.0.8)
Requirement already satisfied: matplotlib>=2.1.0 in /usr/local/lib/python3.10/dist-packages (from pycocotools) (3.7.1)
Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from pycocotools) (1.26.4)
Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (1.3.0)
Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (0.12.1)
Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (4.53.1)
Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (1.4.7)
Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (24.1)
Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (10.4.0)
Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (3.1.4)
Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.1.0->pycocotools) (2.8.2)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib>=2.1.0->pycocotools) (1.16.0)

!wget https://raw.githubusercontent.com/pytorch/vision/main/references/detection/engine.py
!wget https://raw.githubusercontent.com/pytorch/vision/main/references/detection/utils.py
!wget https://raw.githubusercontent.com/pytorch/vision/main/references/detection/coco_utils.py
!wget https://raw.githubusercontent.com/pytorch/vision/main/references/detection/coco_eval.py
!wget https://raw.githubusercontent.com/pytorch/vision/main/references/detection/transforms.py
    
from engine import evaluate

--2024-12-25 08:30:18--  https://raw.githubusercontent.com/pytorch/vision/main/references/detection/engine.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.110.133, 185.199.109.133, 185.199.108.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.110.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 4063 (4.0K) [text/plain]
Saving to: ‘engine.py’

engine.py           100%[===================>]   3.97K  --.-KB/s    in 0s      

2024-12-25 08:30:18 (44.1 MB/s) - ‘engine.py’ saved [4063/4063]

--2024-12-25 08:30:18--  https://raw.githubusercontent.com/pytorch/vision/main/references/detection/utils.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.111.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 8388 (8.2K) [text/plain]
Saving to: ‘utils.py’

utils.py            100%[===================>]   8.19K  --.-KB/s    in 0s      

2024-12-25 08:30:18 (73.1 MB/s) - ‘utils.py’ saved [8388/8388]

--2024-12-25 08:30:18--  https://raw.githubusercontent.com/pytorch/vision/main/references/detection/coco_utils.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.109.133, 185.199.110.133, 185.199.111.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.109.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 8397 (8.2K) [text/plain]
Saving to: ‘coco_utils.py’

coco_utils.py       100%[===================>]   8.20K  --.-KB/s    in 0s      

2024-12-25 08:30:18 (59.4 MB/s) - ‘coco_utils.py’ saved [8397/8397]

--2024-12-25 08:30:19--  https://raw.githubusercontent.com/pytorch/vision/main/references/detection/coco_eval.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.111.133, 185.199.108.133, 185.199.110.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.111.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6447 (6.3K) [text/plain]
Saving to: ‘coco_eval.py’

coco_eval.py        100%[===================>]   6.30K  --.-KB/s    in 0s      

2024-12-25 08:30:19 (66.9 MB/s) - ‘coco_eval.py’ saved [6447/6447]

--2024-12-25 08:30:19--  https://raw.githubusercontent.com/pytorch/vision/main/references/detection/transforms.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.109.133, 185.199.111.133, 185.199.108.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.109.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 23628 (23K) [text/plain]
Saving to: ‘transforms.py’

transforms.py       100%[===================>]  23.07K  --.-KB/s    in 0.001s  

2024-12-25 08:30:19 (15.8 MB/s) - ‘transforms.py’ saved [23628/23628]

## TODO: Prepare your own information
class config2:
    
    ## roots for training & valid
    root = "/kaggle/working/train"
    info_root = "/kaggle/working"
    save_root = "/kaggle/working"

    ## for test images
    test_root = '/kaggle/working/test'
    info_root_test = '/kaggle/input/hwk05-data/hwk05_data/train'
    
    num_classes = 8 #(for fasterrcnn: background + # of classes): 1+7=8
    
    batch_size = 8
    epochs = 20
    weight_decay = 1e-4
    lr = 0.005
    momentum = 0.9
    seed = 42
    workers = 8
    
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

def seed_everything(seed):
   
    random.seed(seed) # Set Python random seed    
    np.random.seed(seed) # Set NumPy random seed    
    torch.manual_seed(seed) # Set PyTorch random seed for CPU and GPU
    if torch.cuda.is_available():
        torch.cuda.manual_seed(seed)
        torch.cuda.manual_seed_all(seed)
    
    # Set PyTorch deterministic operations for cudnn backend
    torch.backends.cudnn.deterministic = True
    torch.backends.cudnn.benchmark = False

    
seed_everything(config2.seed)

## TO DO: Augmentation?
class medTransform:
    def __init__(self, train=False):    
        if train:
            self.transforms = v2.Compose(
                [
                    v2.ToImage(),  # Used while using PIL image
                    v2.RandomRotation(degrees=5),  
                    v2.ToDtype(torch.float32, scale=True),
                ]
            )
        else:
            self.transforms = v2.Compose(
                [
                    v2.ToImage(),  # Used while using PIL image
                    v2.ToDtype(torch.float32, scale=True),
                ]
            )       

    def __call__(self, x, bboxes):
        if isinstance(x, torch.Tensor):
            height, width = x.shape[-2], x.shape[-1]  # (C, H, W) format
        else:  # Assuming x is a PIL Image
            width, height = x.size
        # Loading format is COCO bboxes[x,y,w,h]
        bboxes = tv_tensors.BoundingBoxes(bboxes, format="XYWH", canvas_size=(height, width))
        return self.transforms(x, bboxes)

class MedDataset(Dataset):
    
    def __init__(self, root, info_root, split, transforms = None):
        self.split = split
        self.root = root
        self.info_root = info_root
        self.transforms = transforms
        self.coco = COCO(os.path.join(self.info_root, f"{self.split}.json"))
        self.ids = list(sorted(self.coco.imgs.keys()))
    
    def get_image(self, img_id: int):
        image_path = os.path.join(self.root,self.coco.loadImgs(img_id)[0]['file_name'])
        image = Image.open(image_path).convert("RGB")
        return image

    def get_annotation(self, img_id: int):        
        return self.coco.loadAnns(self.coco.getAnnIds(img_id))
          
    def __getitem__(self, index):
        normal = False
        img_id = self.ids[index]
        image = self.get_image(img_id) 
        annotation = self.get_annotation(img_id)
        
        bboxes = [a['bbox']  for a in annotation]
        category_ids = [a['category_id']  for a in annotation]
        if bboxes == []:
            normal = True     
            
        if self.transforms:
            image, bboxes = self.transforms(image, bboxes)     
            
        def reformat_bboxes(boxes):
            return [[val[0], val[1], val[0] + val[2], val[1] + val[3]] for val in boxes]
        
        if normal != True:
            ## Recall that the original format is COCO
            bboxes = reformat_bboxes(bboxes)   

        def create_target(bboxes, normal):
            if normal:
                return {
                    'boxes': torch.zeros((0, 4), dtype=torch.float32),  # Empty boxes
                    'labels': torch.tensor(category_ids, dtype=torch.int64),
                    'image_id': img_id,
                    'area': torch.zeros((0,), dtype=torch.float32),  # Empty areas
                    'iscrowd': torch.zeros((0,), dtype=torch.int64),  # Empty tensor for iscrowd
                }
            else:
            
                return {
                    'boxes': torch.tensor(bboxes),
                    'labels': torch.tensor(category_ids, dtype=torch.int64),
                    'image_id': img_id,
                    'area': torch.tensor([(bbox[2] - bbox[0]) * (bbox[3] - bbox[1]) for bbox in bboxes], dtype=torch.float32),
                    'iscrowd': torch.tensor([a['iscrowd'] for a in annotation], dtype=torch.int64)
                }
       
        targets = create_target(bboxes,normal)
        return image, targets
        
    def __len__(self):
        return len(self.ids)

def collate_fn(batch: list[torch.tensor, dict]):
    return tuple(zip(*batch))

def plot_image_with_boxes(image_tensor, boxes_dict):
    image_np = image_tensor.permute(1, 2, 0).numpy()
    fig, ax = plt.subplots(1)
    # Display the image
    ax.imshow(image_np)
    for box in boxes_dict['boxes']:
        # Extract coordinates (x0, y0, x1, y1)
        x0, y0, x1, y1 = box
        # Calculate the height as (y0 - y1) since y0 is the top and y1 is the bottom
        height = y1 - y0
        # Create a rectangle patch with (x0, y0) as the top-left corner
        rect = patches.Rectangle((x0, y0), x1 - x0, height, linewidth=2, edgecolor='r', facecolor='none')
        ax.add_patch(rect)
    plt.show()

def fasterrcnn(num_classes):
    model = models.detection.fasterrcnn_resnet50_fpn(weights='COCO_V1') 
    in_features = model.roi_heads.box_predictor.cls_score.in_features
    model.roi_heads.box_predictor = None
    model.roi_heads.box_predictor = FastRCNNPredictor(in_features, num_classes)
    
    return model

def train_one_epoch(model, train_loader, optimizer, epoch, device):
    model.train()
    
    train_loss = []
    train_loss_dict = []
    
    lr_scheduler = None

    for images, targets in tqdm(train_loader):
        images = [image.to(device) for image in images]
        targets = [{k: (torch.tensor(v,device=device) if not isinstance(v, torch.Tensor) else v.to(device)) for k, v in t.items()} for t in targets]

        loss_dict = model(images, targets)
        losses = sum(loss for loss in loss_dict.values())

        batch_loss_value = losses.item()
        batch_loss_dict = {k: v.item() for k, v in loss_dict.items()}

        train_loss.append(batch_loss_value)
        train_loss_dict.append(batch_loss_dict)
      
        optimizer.zero_grad()
        losses.backward()
        optimizer.step()
    
        if lr_scheduler is not None:
            lr_scheduler.step()
        
    train_loss = np.mean(train_loss)    
    train_loss_dict = pd.DataFrame(train_loss_dict).mean()
    train_loss_classifier = train_loss_dict.loss_classifier
    train_loss_box_reg = train_loss_dict.loss_box_reg
    train_loss_rpn_box_reg = train_loss_dict.loss_rpn_box_reg
    train_loss_objectness = train_loss_dict.loss_objectness

    return train_loss, train_loss_classifier, train_loss_box_reg, train_loss_rpn_box_reg, train_loss_objectness

def validation(model, val_loader, device):
    model.train()
    #model.eval()
    for m in model.modules():
        if isinstance(m, torchvision.ops.Conv2dNormActivation):
            m.eval()
        if isinstance(m, torchvision.ops.FrozenBatchNorm2d):
            m.eval()
        if isinstance(m, torch.nn.BatchNorm2d):
            m.eval()
    val_loss = []
    val_loss_dict = []
    with torch.no_grad():
        for images, targets in tqdm(val_loader):
            images = [image.to(device) for image in images]
            targets = [{k: (torch.tensor(v,device=device) if not isinstance(v, torch.Tensor) else v.to(device)) for k, v in t.items()} for t in targets]

            loss = model(images, targets)
            total_loss = sum(l for l in loss.values())
            
            loss_value = total_loss.item()
            loss_dict = {k: v.item() for k, v in loss.items()}
            
            val_loss.append(loss_value)
            val_loss_dict.append(loss_dict)
    
    val_loss = np.mean(val_loss)
    
    val_loss_dict = pd.DataFrame(val_loss_dict).mean()
    val_loss_classifier = val_loss_dict.loss_classifier
    val_loss_box_reg = val_loss_dict.loss_box_reg
    val_loss_rpn_box_reg = val_loss_dict.loss_rpn_box_reg
    val_loss_objectness = val_loss_dict.loss_objectness
    
    return val_loss, val_loss_classifier, val_loss_box_reg, val_loss_rpn_box_reg, val_loss_objectness

def seed_worker(worker_id):
    worker_seed = torch.initial_seed() % 2**32
    np.random.seed(worker_seed)
    random.seed(worker_seed)

def main():
    
    seed_everything(config2.seed)    
    g = torch.Generator()
    g.manual_seed(config2.seed)
    
    train_dataset = MedDataset(root = config2.root, info_root = config2.info_root, split = "train", transforms = medTransform(train=True))
    val_dataset = MedDataset(root = config2.root, info_root = config2.info_root, split = "val",  transforms = medTransform(train=False))

    train_loader = DataLoader(train_dataset, batch_size = config2.batch_size, shuffle = True,
                              num_workers=config2.workers, collate_fn = collate_fn,pin_memory=True
                             )
    val_loader = DataLoader(val_dataset, batch_size = config2.batch_size, shuffle = False,
                            num_workers=config2.workers, worker_init_fn=seed_worker,
                            generator=g, collate_fn = collate_fn,pin_memory=True
                           )

    
    device = config2.device
    model =  fasterrcnn(num_classes = config2.num_classes)
    model.to(device)
    
    parameters = [p for p in model.parameters() if p.requires_grad]

    optimizer = torch.optim.SGD(parameters, lr = config2.lr, momentum = config2.momentum, nesterov = True, weight_decay = config2.weight_decay)
    best_val_loss = float("inf")
    best_map50 = 0.0
    history = {
        "train": {
            "loss": [],
            "loss_classifier": [],
            "loss_box_reg": [],
            "loss_rpn_box_reg": [],
            "loss_objectness": []
        },
        "val": {
            "loss": [],
            "loss_classifier": [],
            "loss_box_reg": [],
            "loss_rpn_box_reg": [],
            "loss_objectness": []
        },
        "map50":{
            "train":[],
            "valid":[],
        }
    }
    best_idx = 0
    print('start')
    for epoch in range(config2.epochs):
        print()
        train_loss, train_loss_classifier, train_loss_box_reg, train_loss_rpn_box_reg, train_loss_objectness = train_one_epoch(
            model, train_loader, optimizer, epoch, device,
        )
        
        val_loss, val_loss_classifier, val_loss_box_reg, val_loss_rpn_box_reg, val_loss_objectness = validation(
            model, val_loader, device
        )

        ## Training
        history["train"]["loss"].append(train_loss)
        history["train"]["loss_classifier"].append(train_loss_classifier)
        history["train"]["loss_box_reg"].append(train_loss_box_reg)
        history["train"]["loss_rpn_box_reg"].append(train_loss_rpn_box_reg)
        history["train"]["loss_objectness"].append(train_loss_objectness)
        ## Validation
        history["val"]["loss"].append(val_loss)
        history["val"]["loss_classifier"].append(val_loss_classifier)
        history["val"]["loss_box_reg"].append(val_loss_box_reg)
        history["val"]["loss_rpn_box_reg"].append(val_loss_rpn_box_reg)
        history["val"]["loss_objectness"].append(val_loss_objectness)

        
        print(f'Epoch: {epoch+1}/{config2.epochs} | LR: {optimizer.state_dict()["param_groups"][0]["lr"]:.6f}')

        print("*****Training*****")
        print(f'Loss: {train_loss:.4f} | Classifier Loss: {train_loss_classifier:.4f} | Box Reg Loss: {train_loss_box_reg:.4f} | RPN Box Reg Loss: {train_loss_rpn_box_reg:.4f} | Objectness Loss: {train_loss_objectness:.4f}')
        train_evaluator = evaluate(model, train_loader, device = device)
        print("*****Validation*****")
        print(f'Loss: {val_loss:.4f} | Classifier Loss: {val_loss_classifier:.4f} | Box Reg Loss: {val_loss_box_reg:.4f} | RPN Box Reg Loss: {val_loss_rpn_box_reg:.4f} | Objectness Loss: {val_loss_objectness:.4f}')
        valid_evaluator = evaluate(model, val_loader, device = device)
        
        train_map50 = train_evaluator.coco_eval['bbox'].stats[1]
        valid_map50 = valid_evaluator.coco_eval['bbox'].stats[1]
        
        history["map50"]["train"].append(train_map50)   
        history["map50"]["valid"].append(valid_map50)        

        ## TODO save your model

        if valid_map50  > best_map50:
            best_map50 = valid_map50
            save_file = {
                "model": model.state_dict(),
                "optimizer": optimizer.state_dict(),
                "epoch": epoch,
                "args": config2
            }
            best_idx=epoch
            
            torch.save(save_file, os.path.join(config2.save_root,"final.pth"))
        
    print(f'Best epoch in {best_idx+1}') 


    ## Evaluation result 
    plt.figure(figsize = (12, 5))
    plt.subplot(1, 2, 1)
    plt.plot(range(config2.epochs), history["map50"]["train"], label = 'Training map50')
    plt.plot(range(config2.epochs), history["map50"]["valid"], label = 'Validation map50')
    plt.xlabel('Epoch')
    plt.ylabel('map')
    plt.legend()
    plt.title('Validation and Testing map50')
    plt.show()
    
    plt.figure(figsize = (12, 5))
    plt.subplot(1, 2, 1)
    plt.plot(range(config2.epochs), history["train"]["loss"], label = 'Training Loss')
    plt.plot(range(config2.epochs), history["val"]["loss"], label = 'Validation Loss')
    plt.xlabel('Epoch')
    plt.ylabel('Loss')
    plt.legend()
    plt.title('Training and Validation Loss Curves')
    plt.show()
        
    plt.figure(figsize = (12, 5))
    plt.subplot(1, 2, 1)
    plt.plot(range(config2.epochs), history["train"]["loss_classifier"], label = 'Training Classifier Loss')
    plt.plot(range(config2.epochs), history["val"]["loss_classifier"], label = 'Validation Classifier Loss')
    plt.xlabel('Epoch')
    plt.ylabel('Classifier Loss')
    plt.legend()
    plt.title('Training and Validation Classifier Loss Curves')
    plt.show()
        
    plt.figure(figsize = (12, 5))
    plt.subplot(1, 2, 1)
    plt.plot(range(config2.epochs), history["train"]["loss_box_reg"], label = 'Training Box Reg Loss')
    plt.plot(range(config2.epochs), history["val"]["loss_box_reg"], label = 'Validation Box Reg Loss')
    plt.xlabel('Epoch')
    plt.ylabel('Box Reg Loss')
    plt.legend()
    plt.title('Training and Validation Box Reg Loss Curves')
    plt.show()
        
    plt.figure(figsize = (12, 5))
    plt.subplot(1, 2, 1)
    plt.plot(range(config2.epochs), history["train"]["loss_rpn_box_reg"], label = 'Training RPN Box Reg Loss')
    plt.plot(range(config2.epochs), history["val"]["loss_rpn_box_reg"], label = 'Validation RPN Box Reg Loss')
    
    plt.xlabel('Epoch')
    plt.ylabel('RPN Box Reg Loss')
    plt.legend()
    plt.title('Training and Validation RPN Box Reg Loss Curves')
    plt.show()
        
    plt.figure(figsize = (12, 5))
    plt.subplot(1, 2, 1)
    plt.plot(range(config2.epochs), history["train"]["loss_objectness"], label = 'Training Objectness Loss')
    plt.plot(range(config2.epochs), history["val"]["loss_objectness"], label = 'Validation Objectness Loss')

    plt.xlabel('Epoch')
    plt.ylabel('Objectness Loss')
    plt.legend()
    plt.title('Training and Validation Objectness Loss Curves')
    plt.show()

## IMAGENET 3
if __name__ == "__main__":
    main()

loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!

Downloading: "https://download.pytorch.org/models/fasterrcnn_resnet50_fpn_coco-258fb6c6.pth" to /root/.cache/torch/hub/checkpoints/fasterrcnn_resnet50_fpn_coco-258fb6c6.pth
100%|██████████| 160M/160M [00:00<00:00, 197MB/s]

start

  0%|          | 0/35 [00:00<?, ?it/s]

  0%|          | 0/10 [00:00<?, ?it/s]

Epoch: 1/20 | LR: 0.005000
*****Training*****
Loss: 0.3138 | Classifier Loss: 0.1968 | Box Reg Loss: 0.0935 | RPN Box Reg Loss: 0.0089 | Objectness Loss: 0.0145
creating index...
index created!
Test:  [ 0/35]  eta: 0:12:50  model_time: 1.2099 (1.2099)  evaluator_time: 0.1441 (0.1441)  time: 22.0015  data: 20.5881  max mem: 11900
Test:  [34/35]  eta: 0:00:02  model_time: 0.8543 (0.8925)  evaluator_time: 0.0360 (0.0686)  time: 1.9947  data: 1.0389  max mem: 11900
Test: Total time: 0:01:30 (2.5810 s / it)
Averaged stats: model_time: 0.8543 (0.8925)  evaluator_time: 0.0360 (0.0686)
Accumulating evaluation results...
DONE (t=0.19s).
IoU metric: bbox
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.033
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.100
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.015
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.033
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.119
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.269
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.269
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.269
*****Validation*****
Loss: 0.1989 | Classifier Loss: 0.0990 | Box Reg Loss: 0.0825 | RPN Box Reg Loss: 0.0065 | Objectness Loss: 0.0109
creating index...
index created!
Test:  [ 0/10]  eta: 0:01:21  model_time: 0.8241 (0.8241)  evaluator_time: 0.0099 (0.0099)  time: 8.1230  data: 7.2376  max mem: 11900
Test:  [ 9/10]  eta: 0:00:01  model_time: 0.8241 (0.7459)  evaluator_time: 0.0136 (0.0122)  time: 1.5261  data: 0.7240  max mem: 11900
Test: Total time: 0:00:15 (1.5365 s / it)
Averaged stats: model_time: 0.8241 (0.7459)  evaluator_time: 0.0136 (0.0122)
Accumulating evaluation results...
DONE (t=0.06s).
IoU metric: bbox
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.024
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.082
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.009
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.024
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.103
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.259
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.262
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.262

  0%|          | 0/35 [00:00<?, ?it/s]

  0%|          | 0/10 [00:00<?, ?it/s]

Epoch: 2/20 | LR: 0.005000
*****Training*****
Loss: 0.2156 | Classifier Loss: 0.1035 | Box Reg Loss: 0.0991 | RPN Box Reg Loss: 0.0068 | Objectness Loss: 0.0062
creating index...
index created!
Test:  [ 0/35]  eta: 0:13:27  model_time: 0.9888 (0.9888)  evaluator_time: 0.0900 (0.0900)  time: 23.0640  data: 21.9080  max mem: 11900
Test:  [34/35]  eta: 0:00:02  model_time: 0.8700 (0.8832)  evaluator_time: 0.0192 (0.0543)  time: 1.7739  data: 0.8249  max mem: 11900
Test: Total time: 0:01:27 (2.5023 s / it)
Averaged stats: model_time: 0.8700 (0.8832)  evaluator_time: 0.0192 (0.0543)
Accumulating evaluation results...
DONE (t=0.16s).
IoU metric: bbox
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.093
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.228
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.057
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.093
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.298
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.357
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.359
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.359
*****Validation*****
Loss: 0.1860 | Classifier Loss: 0.0901 | Box Reg Loss: 0.0830 | RPN Box Reg Loss: 0.0073 | Objectness Loss: 0.0056
creating index...
index created!
Test:  [ 0/10]  eta: 0:01:20  model_time: 0.8165 (0.8165)  evaluator_time: 0.0125 (0.0125)  time: 8.0725  data: 7.1909  max mem: 11900
Test:  [ 9/10]  eta: 0:00:01  model_time: 0.8165 (0.7453)  evaluator_time: 0.0123 (0.0119)  time: 1.5206  data: 0.7194  max mem: 11900
Test: Total time: 0:00:15 (1.5344 s / it)
Averaged stats: model_time: 0.8165 (0.7453)  evaluator_time: 0.0123 (0.0119)
Accumulating evaluation results...
DONE (t=0.06s).
IoU metric: bbox
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.085
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.204
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.050
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.085
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.269
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.311
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.311
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.311

  0%|          | 0/35 [00:00<?, ?it/s]

import os
import pandas as pd
import torch
from torchvision.transforms import functional as F
from tqdm.notebook import tqdm
from PIL import Image

# 定義測試數據集類
class TestDataset:
    def __init__(self, root, info_root, transforms=None):
        self.root = root
        self.transforms = transforms
        self.test_info = pd.read_csv(os.path.join(info_root, "test.csv"))

    def __len__(self):
        return len(self.test_info)

    def __getitem__(self, idx):
        # 使用 Filename 來定位文件
        file_name = self.test_info.iloc[idx]["Filename"]
        img_path = os.path.join(self.root, "image", file_name.replace(".dcm", ".jpg"))  # 確保對應到 .jpg
        if not os.path.exists(img_path):
            raise FileNotFoundError(f"Image file not found: {img_path}")
        image = Image.open(img_path).convert("RGB")
        width = self.test_info.iloc[idx]["Width"]
        height = self.test_info.iloc[idx]["Height"]
        if self.transforms:
            image = self.transforms(image)
        return self.test_info.iloc[idx]["ID"], image, width, height

# 測試推論函數
def inference_on_test_set(model, test_loader, device):
    model.eval()
    results = []
    with torch.no_grad():
        for img_ids, images, widths, heights in tqdm(test_loader):
            images = [img.to(device) for img in images]
            outputs = model(images)
            for img_id, output, img, w, h in zip(img_ids, outputs, images, widths, heights):
                w, h = float(w), float(h)  # 將寬度和高度轉為浮點數
                for box, label, score in zip(output["boxes"], output["labels"], output["scores"]):
                    if score < 0.5:  # 過濾低置信度框
                        continue
                    # 邊界框歸一化處理
                    xmin, ymin, xmax, ymax = box.tolist()
                    xmin, xmax = xmin / w, xmax / w
                    ymin, ymax = ymin / h, ymax / h
                    category = config2.categories[label.item() - 1]  # 從類別索引轉換為名稱
                    results.append({
                        "ID": img_id,
                        "category": category,
                        "score": score.item(),
                        "xmin": xmin,
                        "xmax": xmax,
                        "ymin": ymin,
                        "ymax": ymax
                    })
    return results

# 保存結果到 CSV
def save_results_to_csv(results, output_path):
    df = pd.DataFrame(results)
    df.to_csv(output_path, index=False)
    print(f"Results saved to {output_path}")
    return df

# 主函數
def main():
    # 設置模型設備
    device = config2.device

    # 初始化模型
    model = fasterrcnn(num_classes=config2.num_classes)
    model.load_state_dict(torch.load(os.path.join(config2.save_root, "final.pth"))["model"])
    model.to(device)

    # 初始化測試數據集
    test_dataset = TestDataset(
        root=config2.test_root, 
        info_root=config2.info_root_test,
        transforms=F.to_tensor
    )
    test_loader = torch.utils.data.DataLoader(
        test_dataset,
        batch_size=1,  # 一次處理一張影像
        shuffle=False,
        collate_fn=lambda x: tuple(zip(*x))
    )

    # 進行推論
    print("Running inference on test set...")
    results = inference_on_test_set(model, test_loader, device)

    # 保存結果
    df = save_results_to_csv(results, os.path.join(config2.save_root, "submission.csv"))
    return df

if __name__ == "__main__":
    # 在 config2 中定義類別名稱
    class config2:
        test_root = "/kaggle/working/test"
        info_root_test = "/kaggle/input/hwk05-data/hwk05_data/test"
        save_root = "/kaggle/working"
        device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
        num_classes = 8
        categories = [
            "aortic_curvature",
            "aortic_atherosclerosis_calcification",
            "cardiac_hypertrophy",
            "intercostal_pleural_thickening",
            "lung_field_infiltration",
            "degenerative_joint_disease_of_the_thoracic",
            "scoliosis",
            "normal"
        ]
    pred_df=main()

Running inference on test set...

  0%|          | 0/113 [00:00<?, ?it/s]

Results saved to /kaggle/working/submission.csv

# 顯示前10行、後10行和數據形狀
print("First 10 rows:")
print(pred_df.head(10))
print("\nLast 10 rows:")
print(pred_df.tail(10))
print("\nDataframe shape:")
print(pred_df.shape)

First 10 rows:
                      ID                                    category  \
0  TDR02_20161123_145314        aortic_atherosclerosis_calcification   
1  TDR02_20161123_145314                            aortic_curvature   
2  TDR02_20161123_145314                     lung_field_infiltration   
3  TDR02_20161123_145314  degenerative_joint_disease_of_the_thoracic   
4  TDR01_20171106_111727  degenerative_joint_disease_of_the_thoracic   
5  TDR01_20171106_111727                     lung_field_infiltration   
6  TDR01_20171106_111727                            aortic_curvature   
7  TDR01_20171106_111727                                   scoliosis   
8  TDR01_20180510_090210  degenerative_joint_disease_of_the_thoracic   
9  TDR01_20180511_092549                                   scoliosis   

      score      xmin      xmax      ymin      ymax  
0  0.961782  0.501516  0.664308  0.200134  0.381468  
1  0.787029  0.374931  0.675504  0.232649  0.719997  
2  0.758716  0.063560  0.959290  0.010014  0.866449  
3  0.549290  0.365644  0.673442  0.016905  0.774482  
4  0.837700  0.360594  0.624836  0.072813  0.807553  
5  0.596343  0.053755  0.940258  0.030877  0.854882  
6  0.542351  0.359334  0.641187  0.229484  0.678211  
7  0.519861  0.407437  0.616236  0.185018  0.890281  
8  0.738120  0.363038  0.617012  0.122861  0.809470  
9  0.724835  0.392327  0.619073  0.210029  0.839867  

Last 10 rows:
                        ID                                    category  \
266  TDR02_20161125_122319  degenerative_joint_disease_of_the_thoracic   
267  TDR02_20161125_122319                                   scoliosis   
268  TDR02_20180123_115426  degenerative_joint_disease_of_the_thoracic   
269  TDR02_20180123_115426                     lung_field_infiltration   
270  TDR01_20180508_173616                                   scoliosis   
271  TDR02_20161118_145330        aortic_atherosclerosis_calcification   
272  TDR02_20161118_145330                         cardiac_hypertrophy   
273  TDR02_20161118_145330                     lung_field_infiltration   
274  TDR02_20161118_145330                            aortic_curvature   
275  TDR02_20161118_145330  degenerative_joint_disease_of_the_thoracic   

        score      xmin      xmax      ymin      ymax  
266  0.699098  0.381842  0.635729  0.054117  0.807661  
267  0.500009  0.405378  0.613517  0.294539  0.858112  
268  0.781155  0.365377  0.629589  0.083213  0.827556  
269  0.720102  0.055466  0.915525  0.037781  0.866403  
270  0.661993  0.388692  0.606562  0.226665  0.805151  
271  0.979094  0.419940  0.596462  0.245768  0.422824  
272  0.934155  0.297666  0.760472  0.360590  0.693891  
273  0.766747  0.055702  0.894140  0.056281  0.816584  
274  0.668177  0.273974  0.609571  0.255313  0.668751  
275  0.504318  0.283422  0.588073  0.074076  0.742070  

Dataframe shape:
(276, 7)

# import libraries

# basic
import warnings
warnings.filterwarnings('ignore')

import os
import random
import numpy as np
import pandas as pd

# visualization
import cv2
from PIL import Image
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
!pip install grad-cam
import pytorch_grad_cam
from pytorch_grad_cam import EigenCAM, AblationCAM
from pytorch_grad_cam.ablation_layer import AblationLayerFasterRCNN
from pytorch_grad_cam.utils.model_targets import FasterRCNNBoxScoreTarget
from pytorch_grad_cam.utils.reshape_transforms import fasterrcnn_reshape_transform
from pytorch_grad_cam.utils.image import show_cam_on_image

# PyTorch
import torch
import torchvision
from torchvision import models
from torchvision.models.detection.faster_rcnn import FastRCNNPredictor
from torchvision.transforms import v2
from torchvision.transforms.v2 import functional as F

Collecting grad-cam
  Downloading grad-cam-1.5.4.tar.gz (7.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 7.8/7.8 MB 58.0 MB/s eta 0:00:0000:0100:01
  Installing build dependencies ... done
  Getting requirements to build wheel ... done
  Preparing metadata (pyproject.toml) ... done
Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from grad-cam) (1.26.4)
Requirement already satisfied: Pillow in /usr/local/lib/python3.10/dist-packages (from grad-cam) (10.4.0)
Requirement already satisfied: torch>=1.7.1 in /usr/local/lib/python3.10/dist-packages (from grad-cam) (2.4.1+cu121)
Requirement already satisfied: torchvision>=0.8.2 in /usr/local/lib/python3.10/dist-packages (from grad-cam) (0.19.1+cu121)
Collecting ttach (from grad-cam)
  Downloading ttach-0.0.3-py3-none-any.whl.metadata (5.2 kB)
Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from grad-cam) (4.66.5)
Requirement already satisfied: opencv-python in /usr/local/lib/python3.10/dist-packages (from grad-cam) (4.10.0.84)
Requirement already satisfied: matplotlib in /usr/local/lib/python3.10/dist-packages (from grad-cam) (3.7.1)
Requirement already satisfied: scikit-learn in /usr/local/lib/python3.10/dist-packages (from grad-cam) (1.2.2)
Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch>=1.7.1->grad-cam) (3.16.1)
Requirement already satisfied: typing-extensions>=4.8.0 in /usr/local/lib/python3.10/dist-packages (from torch>=1.7.1->grad-cam) (4.12.2)
Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch>=1.7.1->grad-cam) (1.13.3)
Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.7.1->grad-cam) (3.3)
Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.7.1->grad-cam) (3.1.4)
Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch>=1.7.1->grad-cam) (2024.6.1)
Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (1.3.0)
Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (0.12.1)
Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (4.53.1)
Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (1.4.7)
Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (24.1)
Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (3.1.4)
Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib->grad-cam) (2.8.2)
Requirement already satisfied: scipy>=1.3.2 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->grad-cam) (1.13.1)
Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->grad-cam) (1.4.2)
Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->grad-cam) (3.5.0)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib->grad-cam) (1.16.0)
Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.7.1->grad-cam) (2.1.5)
Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy->torch>=1.7.1->grad-cam) (1.3.0)
Downloading ttach-0.0.3-py3-none-any.whl (9.8 kB)
Building wheels for collected packages: grad-cam
  Building wheel for grad-cam (pyproject.toml) ... done
  Created wheel for grad-cam: filename=grad_cam-1.5.4-py3-none-any.whl size=39588 sha256=666f98426457c54f6518315f1acf3e1bf652653fc49f931a0dab41fd1f26a479
  Stored in directory: /root/.cache/pip/wheels/50/b0/82/1f97b5348c7fe9f0ce0ba18497202cafa5dec4562bd5292680
Successfully built grad-cam
Installing collected packages: ttach, grad-cam
Successfully installed grad-cam-1.5.4 ttach-0.0.3

class config3:
    
    root = "/kaggle/working/train"
    num_classes = 8
    categories = ['normal', 'aortic_curvature', 'aortic_atherosclerosis_calcification', 
                  'cardiac_hypertrophy', 'intercostal_pleural_thickening', 'lung_field_infiltration', 
                  'degenerative_joint_disease_of_the_thoracic_spine', 'scoliosis']
    seed = 42
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

def seed_everything(seed):
    # Set Python random seed
    random.seed(seed)
    
    # Set NumPy random seed
    np.random.seed(seed)
    
    # Set PyTorch random seed for CPU and GPU
    torch.manual_seed(seed)
    if torch.cuda.is_available():
        torch.cuda.manual_seed_all(seed)
    
    # Set PyTorch deterministic operations for cudnn backend
    torch.backends.cudnn.deterministic = True
    torch.backends.cudnn.benchmark = True
    
seed_everything(config3.seed)

def predict(input_tensor, model, device, detection_threshold):
    outputs = model(input_tensor)
    pred_classes = [config3.categories[i] for i in outputs[0]['labels'].cpu().numpy()]
    pred_labels = outputs[0]['labels'].cpu().numpy()
    pred_scores = outputs[0]['scores'].detach().cpu().numpy()
    pred_bboxes = outputs[0]['boxes'].detach().cpu().numpy()
    
    boxes, classes, labels, indices,scores = [], [], [], [], []
    for index in range(len(pred_scores)):
        if pred_scores[index] >= detection_threshold:
            boxes.append(pred_bboxes[index].astype(np.int32))
            classes.append(pred_classes[index])
            labels.append(pred_labels[index])
            indices.append(index)
            scores.append(pred_scores[index])
    boxes = np.int32(boxes)
    
    return boxes, classes, labels, indices, scores

COLORS = np.random.uniform(0, 255, size=(len(config3.categories), 3))

def draw_boxes(boxes, labels, classes, image):
    for i, box in enumerate(boxes):
        # Convert RGB to BGR for OpenCV
        color = COLORS[labels[i]].astype(int)[::-1]
        
        # Draw the bounding box
        cv2.rectangle(
            image,
            (int(box[0]), int(box[1])),
            (int(box[2]), int(box[3])),
            color.tolist(), 8
        )
        
        # Increase font size and thickness for label
        font_scale = 4 # Increase this value for larger font
        thickness = 10     # Increase thickness for better visibility
        
        # Add class label as text
        cv2.putText(image, classes[i], 
                    (int(box[0]), int(box[1]) - 10),  # Adjust text position
                    cv2.FONT_HERSHEY_SIMPLEX, 
                    font_scale, 
                    color.tolist(), 
                    thickness,
                    lineType=cv2.LINE_AA)
    return image

def get_transform():
    
    transform = v2.Compose(
                [
                    v2.ToImage(), ## Used while using PIL image
                    #v2.ConvertBoundingBoxFormat(tv_tensors.BoundingBoxFormat.XYXY),
                    v2.ToDtype(torch.float32, scale=True),
                    
                ])       
    
    return transform

def plot_eigen_cam_images(transforms, model, cat, threshold):
    rows, cols = 4, 2
    fig = plt.figure(figsize=(10, 20))  # Adjust figure size
    grid = plt.GridSpec(rows, cols)
    
    best_ckpt = torch.load("/kaggle/working/final.pth", map_location=config3.device)
    model.load_state_dict(best_ckpt["model"])
    model.eval().to(config3.device)
    target_layers = [model.backbone]
    
    cam = EigenCAM(model, 
                   target_layers, 
                   reshape_transform=fasterrcnn_reshape_transform)
    
    for i in range(rows * cols):
        all_images = os.listdir(os.path.join(config3.root, config3.categories[i]))        
        image_path = os.path.join(config3.root, config3.categories[i], all_images[0])
        image = Image.open(image_path).convert("RGB")   
        input_tensor = transforms(image)
        input_tensor = input_tensor.to(config3.device)
        input_tensor = input_tensor.unsqueeze(0)
        image = np.array(image)
        image_float_np = np.float32(image) / 255
        boxes, classes, labels, indices, scores = predict(input_tensor, model, config3.device, threshold)        
        image = draw_boxes(boxes, labels, classes, image)
        targets = [FasterRCNNBoxScoreTarget(labels=labels, bounding_boxes=boxes)]
        
        grayscale_cam = cam(input_tensor, targets=targets)
        grayscale_cam = grayscale_cam[0, :]
        cam_image = show_cam_on_image(image_float_np, grayscale_cam, use_rgb=True)
        image_with_bounding_boxes = draw_boxes(boxes, labels, classes, cam_image)
        
        categories = fig.add_subplot(grid[i])
        categories.set_axis_off()
        
        gs = gridspec.GridSpecFromSubplotSpec(1, 1, subplot_spec=grid[i])
        
        ax = fig.add_subplot(gs[0])
        ax.imshow(image_with_bounding_boxes)
        ax.set_title(f"{config3.categories[i]}")
        ax.axis("off")
        
    fig.patch.set_facecolor('white')
    fig.suptitle("EigenCAM Images of 8 categories\n", fontweight='bold', size=16)
    fig.tight_layout()
    fig.subplots_adjust(wspace=0.2, hspace=0.4)  # Add extra space between plots

def plot_ablation_cam_images(transforms, model):
    
    rows, cols = 4, 2
    fig = plt.figure(figsize = (10, 20))
    grid = plt.GridSpec(rows, cols)
    
    best_ckpt = torch.load("/kaggle/working/final.pth", map_location = config3.device)
    model.load_state_dict(best_ckpt["model"])
    model.eval().to(config3.device)
    target_layers = [model.backbone]
    
    cam = AblationCAM(model,
                      target_layers,
                      reshape_transform = fasterrcnn_reshape_transform,
                      ablation_layer = AblationLayerFasterRCNN(),
                      ratio_channels_to_ablate = 1.0)
    
    for i in range(rows * cols):
        
        all_images = os.listdir(os.path.join(config3.root, config3.categories[i]))
        image_path = os.path.join(config3.root, config3.categories[i], all_images[0])
        image = Image.open(image_path).convert("RGB")       
        input_tensor = transforms(image)
        input_tensor = input_tensor.to(config3.device)
        input_tensor = input_tensor.unsqueeze(0)
        image = np.array(image)
        image_float_np = np.float32(image) / 255

        boxes, classes, labels, indices, scores = predict(input_tensor, model, config3.device, 0)
        image = draw_boxes(boxes, labels, classes, image)
        targets = [FasterRCNNBoxScoreTarget(labels = labels, bounding_boxes = boxes)]
        
        grayscale_cam = cam(input_tensor, targets = targets)
        grayscale_cam = grayscale_cam[0, :]
        cam_image = show_cam_on_image(image_float_np, grayscale_cam, use_rgb = True)
        image_with_bounding_boxes = draw_boxes(boxes, labels, classes, cam_image)
        
        categories = fig.add_subplot(grid[i])
        categories.set_axis_off()
        
        gs = gridspec.GridSpecFromSubplotSpec(1, 1, subplot_spec = grid[i])
        
        ax = fig.add_subplot(gs[0])
        ax.imshow(image_with_bounding_boxes)
        ax.set_title(f"{config3.categories[i]}")
        ax.axis("off")
        
    fig.patch.set_facecolor('white')
    fig.suptitle("AblationCAM Images of 8 categories\n", fontweight = 'bold', size = 16)
    fig.tight_layout()

result = plot_eigen_cam_images(transforms = get_transform(), model = fasterrcnn(config3.num_classes), cat=0, threshold=0)

plot_ablation_cam_images(transforms = get_transform(), model = fasterrcnn(config3.num_classes))

100%|██████████| 40/40 [00:44<00:00,  1.12s/it]
100%|██████████| 40/40 [00:49<00:00,  1.23s/it]
100%|██████████| 40/40 [00:52<00:00,  1.32s/it]
100%|██████████| 40/40 [00:52<00:00,  1.31s/it]
100%|██████████| 40/40 [00:55<00:00,  1.40s/it]
100%|██████████| 40/40 [00:39<00:00,  1.00it/s]
100%|██████████| 40/40 [00:40<00:00,  1.01s/it]
100%|██████████| 40/40 [00:54<00:00,  1.37s/it]

!zip -r file.zip /kaggle/working

  adding: kaggle/working/ (stored 0%)
  adding: kaggle/working/final.pth (deflated 7%)
  adding: kaggle/working/coco_utils.py (deflated 71%)
  adding: kaggle/working/train.json (deflated 83%)
  adding: kaggle/working/engine.py (deflated 66%)
  adding: kaggle/working/transforms.py (deflated 76%)
  adding: kaggle/working/__pycache__/ (stored 0%)
  adding: kaggle/working/__pycache__/engine.cpython-310.pyc (deflated 44%)
  adding: kaggle/working/__pycache__/transforms.cpython-310.pyc (deflated 54%)
  adding: kaggle/working/__pycache__/coco_eval.cpython-310.pyc (deflated 51%)
  adding: kaggle/working/__pycache__/coco_utils.cpython-310.pyc (deflated 53%)
  adding: kaggle/working/__pycache__/utils.cpython-310.pyc (deflated 50%)
  adding: kaggle/working/train/ (stored 0%)
  adding: kaggle/working/train/aortic_curvature/ (stored 0%)
  adding: kaggle/working/train/aortic_curvature/FILE0.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/220_5.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/220_4.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/FILE0_1.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/220.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/220_1.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/4440_14.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/220_6.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/220_0.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/FILE0_7.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/Volume0_5.jpg (deflated 9%)
  adding: kaggle/working/train/aortic_curvature/Volume0_1.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/10_0.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_curvature/10_7.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/4440_9.jpg (deflated 9%)
  adding: kaggle/working/train/aortic_curvature/FILE0_2.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/4440_1.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/220_2.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/FILE0_3.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/FILE0_6.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/10_6.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/220_7.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/4440_15.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/10_2.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/4440_7.jpg (deflated 9%)
  adding: kaggle/working/train/aortic_curvature/10_10.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/FILE0_4.jpg (deflated 10%)
  adding: kaggle/working/train/aortic_curvature/10_4.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/10_15.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/10_3.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/Volume0_3.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_curvature/10_12.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/220_14.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_curvature/10_13.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_curvature/4440_5.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/4440.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_curvature/FILE0_0.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/220_11.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_curvature/220_10.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/10_5.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/10_8.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/10_11.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_curvature/Volume0.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/FILE0_5.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/220_12.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_curvature/4440_6.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/Volume0_4.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_curvature/4440_12.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_curvature/220_9.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_curvature/10.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_curvature/10_1.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_curvature/4440_8.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/ (stored 0%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_6.jpg (deflated 4%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_0.jpg (deflated 11%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_7.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_10.jpg (deflated 9%)
  adding: kaggle/working/train/lung_field_infiltration/10_0 (2).jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_1.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_13.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/220_1.jpg (deflated 3%)
  adding: kaggle/working/train/lung_field_infiltration/220_0.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_7.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_16.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_1.jpg (deflated 12%)
  adding: kaggle/working/train/lung_field_infiltration/10_7 (2).jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/10_0.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/10_7.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_2.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/4440_1.jpg (deflated 3%)
  adding: kaggle/working/train/lung_field_infiltration/220_2.jpg (deflated 3%)
  adding: kaggle/working/train/lung_field_infiltration/10_9.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_11.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/10_3 (2).jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/10_1 (2).jpg (deflated 4%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_12.jpg (deflated 10%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_3.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_23.jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_17.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_2.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_6.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/10_6.jpg (deflated 4%)
  adding: kaggle/working/train/lung_field_infiltration/10_6 (2).jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/10_2.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/4440_0.jpg (deflated 3%)
  adding: kaggle/working/train/lung_field_infiltration/10_10.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_13.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/10_4.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/10_3.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_3.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/10_12.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/10_13.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_8.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/4440_5.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_16.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/10 (2).jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_0.jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/10_4 (2).jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_18.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/10_2 (2).jpg (deflated 4%)
  adding: kaggle/working/train/lung_field_infiltration/10_5.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/10_8.jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/10_11.jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_5.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/220_3.jpg (deflated 3%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_9.jpg (deflated 7%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_21.jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_20.jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_15.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/10_11 (2).jpg (deflated 5%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_8.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_12.jpg (deflated 11%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_9.jpg (deflated 10%)
  adding: kaggle/working/train/lung_field_infiltration/Volume0_4.jpg (deflated 8%)
  adding: kaggle/working/train/lung_field_infiltration/FILE0_19.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/10_8 (2).jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/4440_4.jpg (deflated 9%)
  adding: kaggle/working/train/lung_field_infiltration/10.jpg (deflated 6%)
  adding: kaggle/working/train/lung_field_infiltration/10_1.jpg (deflated 4%)
  adding: kaggle/working/train/lung_field_infiltration/10_5 (2).jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/ (stored 0%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_9a.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_6a.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440b.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0a.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_8a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_0b.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_5a.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_5d.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_1a.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_8c.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_0d.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_2a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0c.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_3a.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0_7c.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_1d.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_2a.jpg (deflated 9%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10a.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0c.jpg (deflated 11%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_1b.jpg (deflated 9%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0_0c.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_4d.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_0b.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220b.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_0a.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_0 (2)d.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_12a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_3b.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_0d.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_2a.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0b.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10b.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_7a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_5a.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_4a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/A0_0d.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_1a.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440_0a.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220a.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0_6c.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_10a.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_0c.jpg (deflated 10%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_7c.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_1b.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0_11c.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0b.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10d.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_0a.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_3a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220 (2)d.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_1 (2)d.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440d.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0d.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_3a.jpg (deflated 7%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_5c.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_4a.jpg (deflated 3%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0_5c.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_0b.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/A0d.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_3d.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_1b.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0d.jpg (deflated 11%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/220_1b.jpg (deflated 8%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/Volume0_2c.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_2d.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_0b.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10_11a.jpg (deflated 5%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_3c.jpg (deflated 9%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/10 (2)d.jpg (deflated 6%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/4440a.jpg (deflated 4%)
  adding: kaggle/working/train/aortic_atherosclerosis_calcification/FILE0_4c.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/ (stored 0%)
  adding: kaggle/working/train/cardiac_hypertrophy/FILE0.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/Volume0_0.jpg (deflated 11%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_0 (2).jpg (deflated 5%)
  adding: kaggle/working/train/cardiac_hypertrophy/220 (2).jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/FILE0_1.jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/220.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/A0_3.jpg (deflated 5%)
  adding: kaggle/working/train/cardiac_hypertrophy/Volume0_1.jpg (deflated 8%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_0.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_7.jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/FILE0_2.jpg (deflated 9%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_9.jpg (deflated 5%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_3 (2).jpg (deflated 5%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_1 (2).jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/Volume0_2.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/FILE0_6.jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_6.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_2.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_4.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/A0_4.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_12.jpg (deflated 5%)
  adding: kaggle/working/train/cardiac_hypertrophy/A0_0.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_13.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/4440.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/10 (2).jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/FILE0_0.jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/A0_2.jpg (deflated 5%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_5.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_8.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/Volume0.jpg (deflated 6%)
  adding: kaggle/working/train/cardiac_hypertrophy/FILE0_5.jpg (deflated 7%)
  adding: kaggle/working/train/cardiac_hypertrophy/10.jpg (deflated 4%)
  adding: kaggle/working/train/cardiac_hypertrophy/10_1.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/ (stored 0%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220_5.jpg (deflated 3%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220_4.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_0 (2).jpg (deflated 3%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220 (2).jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220.jpg (deflated 7%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220_1.jpg (deflated 7%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220_6.jpg (deflated 3%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0_3.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_0.jpg (deflated 5%)
  adding: kaggle/working/train/intercostal_pleural_thickening/4440_1.jpg (deflated 9%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_3 (2).jpg (deflated 3%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_1 (2).jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/4440_2.jpg (deflated 7%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_2.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/4440_0.jpg (deflated 7%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_4.jpg (deflated 5%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0_5.jpg (deflated 6%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_3.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0_1.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0_0.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/4440.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10 (2).jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_2 (2).jpg (deflated 5%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0_2.jpg (deflated 6%)
  adding: kaggle/working/train/intercostal_pleural_thickening/220_3.jpg (deflated 3%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0_6.jpg (deflated 6%)
  adding: kaggle/working/train/intercostal_pleural_thickening/A0.jpg (deflated 4%)
  adding: kaggle/working/train/intercostal_pleural_thickening/4440_4.jpg (deflated 7%)
  adding: kaggle/working/train/intercostal_pleural_thickening/10_1.jpg (deflated 7%)
  adding: kaggle/working/train/normal/ (stored 0%)
  adding: kaggle/working/train/normal/220_83.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_46.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_93.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_69.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_86.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_55.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_52.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_1.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_88.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_85.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_98.jpg (deflated 7%)
  adding: kaggle/working/train/normal/220_6.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_13.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_58.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_43.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_36.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_16.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_63.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_23.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_2.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_22.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_59.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_91.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_97.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_77.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_7.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_31.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_54.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_45.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_75.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_37.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_25.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_74.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_64.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_29.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_49.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_89.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_53.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_80.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_32.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_78.jpg (deflated 6%)
  adding: kaggle/working/train/normal/220_42.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_92.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_35.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_90.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_8.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_57.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_67.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_50.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_11.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_39.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_10.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_44.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_73.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_84.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_70.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_30.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_17.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_3.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_47.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_21.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_19.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_79.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_87.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_66.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_40.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_94.jpg (deflated 5%)
  adding: kaggle/working/train/normal/220_20.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_82.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_18.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_60.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_28.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_51.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_15.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_38.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_56.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_72.jpg (deflated 4%)
  adding: kaggle/working/train/normal/220_33.jpg (deflated 3%)
  adding: kaggle/working/train/normal/220_81.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/ (stored 0%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/FILE0.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/Volume0_0.jpg (deflated 9%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_5.jpg (deflated 7%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_4.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_0 (2).jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220 (2).jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220.jpg (deflated 7%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_1.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_6.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_0.jpg (deflated 7%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_0 (2).jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_4 (2).jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_7 (2).jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_0.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_7 (2).jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_7.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/4440_1.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_3 (2).jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_1 (2).jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/4440_2.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/4440_3.jpg (deflated 9%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_7.jpg (deflated 9%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_6 (2).jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_6 (2).jpg (deflated 8%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_2.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_2 (2).jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/4440_0.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_10.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/A0_5.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_3.jpg (deflated 7%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/A0_4.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_12.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/A0_0.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_13.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/4440.jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/FILE0_0.jpg (deflated 9%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_4 (2).jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_8.jpg (deflated 7%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_11.jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_2 (2).jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_10.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/A0_2.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_8.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_11.jpg (deflated 5%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/Volume0.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_1 (2).jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_3.jpg (deflated 9%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_11 (2).jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_14.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_12.jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_12 (2).jpg (deflated 7%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/A0_6.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_9.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_5 (2).jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/A0.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/4440_4.jpg (deflated 9%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/220_15.jpg (deflated 3%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10.jpg (deflated 4%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_1.jpg (deflated 6%)
  adding: kaggle/working/train/degenerative_joint_disease_of_the_thoracic_spine/10_5 (2).jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/ (stored 0%)
  adding: kaggle/working/train/scoliosis/220_5.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_16.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/220 (2).jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_29.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/220.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_8.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/220_0 (2).jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/10_18.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/A0_22.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/10_0.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/A0_28.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/10_7.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/4440_1.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/220_2.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_20.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/10_17.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_11.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/4440_2.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/A0_21.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_25.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/4440_3.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/10_6.jpg (deflated 8%)
  adding: kaggle/working/train/scoliosis/220_7.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/10_2.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/4440_0.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/4440_7.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/A0_10.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/10_4.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_17.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/10_15.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_5.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/10_3.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/A0_23.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/A0_4.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/10_16.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_24.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_0.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/A0_26.jpg (deflated 6%)
  adding: kaggle/working/train/scoliosis/10_13.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/4440_5.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/4440.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/10 (2).jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_15.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/220_8.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_2.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/10_5.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/10_8.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_19.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/220_1 (2).jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/A0_12.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/A0_18.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/10_14.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_6.jpg (deflated 4%)
  adding: kaggle/working/train/scoliosis/A0_7.jpg (deflated 5%)
  adding: kaggle/working/train/scoliosis/A0.jpg (deflated 3%)
  adding: kaggle/working/train/scoliosis/4440_4.jpg (deflated 7%)
  adding: kaggle/working/train/scoliosis/10.jpg (deflated 7%)
  adding: kaggle/working/train/scoliosis/A0_14.jpg (deflated 5%)
  adding: kaggle/working/test/ (stored 0%)
  adding: kaggle/working/test/image/ (stored 0%)
  adding: kaggle/working/test/image/103.jpg (deflated 4%)
  adding: kaggle/working/test/image/096.jpg (deflated 7%)
  adding: kaggle/working/test/image/008.jpg (deflated 7%)
  adding: kaggle/working/test/image/095.jpg (deflated 7%)
  adding: kaggle/working/test/image/069.jpg (deflated 6%)
  adding: kaggle/working/test/image/055.jpg (deflated 4%)
  adding: kaggle/working/test/image/016.jpg (deflated 9%)
  adding: kaggle/working/test/image/067.jpg (deflated 5%)
  adding: kaggle/working/test/image/013.jpg (deflated 8%)
  adding: kaggle/working/test/image/099.jpg (deflated 6%)
  adding: kaggle/working/test/image/003.jpg (deflated 5%)
  adding: kaggle/working/test/image/019.jpg (deflated 4%)
  adding: kaggle/working/test/image/091.jpg (deflated 4%)
  adding: kaggle/working/test/image/005.jpg (deflated 4%)
  adding: kaggle/working/test/image/081.jpg (deflated 5%)
  adding: kaggle/working/test/image/011.jpg (deflated 3%)
  adding: kaggle/working/test/image/041.jpg (deflated 6%)
  adding: kaggle/working/test/image/084.jpg (deflated 9%)
  adding: kaggle/working/test/image/089.jpg (deflated 2%)
  adding: kaggle/working/test/image/039.jpg (deflated 4%)
  adding: kaggle/working/test/image/006.jpg (deflated 7%)
  adding: kaggle/working/test/image/083.jpg (deflated 4%)
  adding: kaggle/working/test/image/094.jpg (deflated 5%)
  adding: kaggle/working/test/image/048.jpg (deflated 3%)
  adding: kaggle/working/test/image/075.jpg (deflated 8%)
  adding: kaggle/working/test/image/034.jpg (deflated 4%)
  adding: kaggle/working/test/image/111.jpg (deflated 6%)
  adding: kaggle/working/test/image/051.jpg (deflated 3%)
  adding: kaggle/working/test/image/014.jpg (deflated 8%)
  adding: kaggle/working/test/image/052.jpg (deflated 4%)
  adding: kaggle/working/test/image/010.jpg (deflated 7%)
  adding: kaggle/working/test/image/037.jpg (deflated 6%)
  adding: kaggle/working/test/image/079.jpg (deflated 8%)
  adding: kaggle/working/test/image/049.jpg (deflated 6%)
  adding: kaggle/working/test/image/025.jpg (deflated 8%)
  adding: kaggle/working/test/image/085.jpg (deflated 4%)
  adding: kaggle/working/test/image/065.jpg (deflated 4%)
  adding: kaggle/working/test/image/018.jpg (deflated 4%)
  adding: kaggle/working/test/image/063.jpg (deflated 11%)
  adding: kaggle/working/test/image/001.jpg (deflated 8%)
  adding: kaggle/working/test/image/020.jpg (deflated 4%)
  adding: kaggle/working/test/image/088.jpg (deflated 10%)
  adding: kaggle/working/test/image/007.jpg (deflated 6%)
  adding: kaggle/working/test/image/100.jpg (deflated 6%)
  adding: kaggle/working/test/image/080.jpg (deflated 4%)
  adding: kaggle/working/test/image/078.jpg (deflated 7%)
  adding: kaggle/working/test/image/105.jpg (deflated 8%)
  adding: kaggle/working/test/image/074.jpg (deflated 4%)
  adding: kaggle/working/test/image/098.jpg (deflated 8%)
  adding: kaggle/working/test/image/031.jpg (deflated 13%)
  adding: kaggle/working/test/image/109.jpg (deflated 4%)
  adding: kaggle/working/test/image/104.jpg (deflated 7%)
  adding: kaggle/working/test/image/053.jpg (deflated 4%)
  adding: kaggle/working/test/image/023.jpg (deflated 6%)
  adding: kaggle/working/test/image/045.jpg (deflated 6%)
  adding: kaggle/working/test/image/068.jpg (deflated 4%)
  adding: kaggle/working/test/image/044.jpg (deflated 5%)
  adding: kaggle/working/test/image/073.jpg (deflated 3%)
  adding: kaggle/working/test/image/040.jpg (deflated 3%)
  adding: kaggle/working/test/image/021.jpg (deflated 6%)
  adding: kaggle/working/test/image/101.jpg (deflated 9%)
  adding: kaggle/working/test/image/092.jpg (deflated 7%)
  adding: kaggle/working/test/image/015.jpg (deflated 7%)
  adding: kaggle/working/test/image/004.jpg (deflated 4%)
  adding: kaggle/working/test/image/087.jpg (deflated 3%)
  adding: kaggle/working/test/image/032.jpg (deflated 3%)
  adding: kaggle/working/test/image/072.jpg (deflated 6%)
  adding: kaggle/working/test/image/077.jpg (deflated 3%)
  adding: kaggle/working/test/image/047.jpg (deflated 4%)
  adding: kaggle/working/test/image/057.jpg (deflated 6%)
  adding: kaggle/working/test/image/108.jpg (deflated 5%)
  adding: kaggle/working/test/image/110.jpg (deflated 7%)
  adding: kaggle/working/test/image/046.jpg (deflated 4%)
  adding: kaggle/working/test/image/058.jpg (deflated 7%)
  adding: kaggle/working/test/image/066.jpg (deflated 6%)
  adding: kaggle/working/test/image/024.jpg (deflated 8%)
  adding: kaggle/working/test/image/102.jpg (deflated 7%)
  adding: kaggle/working/test/image/012.jpg (deflated 8%)
  adding: kaggle/working/test/image/026.jpg (deflated 4%)
  adding: kaggle/working/test/image/028.jpg (deflated 4%)
  adding: kaggle/working/test/image/112.jpg (deflated 4%)
  adding: kaggle/working/test/image/002.jpg (deflated 5%)
  adding: kaggle/working/test/image/070.jpg (deflated 4%)
  adding: kaggle/working/test/image/061.jpg (deflated 6%)
  adding: kaggle/working/test/image/076.jpg (deflated 5%)
  adding: kaggle/working/test/image/009.jpg (deflated 6%)
  adding: kaggle/working/test/image/029.jpg (deflated 5%)
  adding: kaggle/working/test/image/022.jpg (deflated 3%)
  adding: kaggle/working/test/image/113.jpg (deflated 7%)
  adding: kaggle/working/test/image/086.jpg (deflated 3%)
  adding: kaggle/working/test/image/056.jpg (deflated 6%)
  adding: kaggle/working/test/image/042.jpg (deflated 5%)
  adding: kaggle/working/test/image/071.jpg (deflated 8%)
  adding: kaggle/working/test/image/027.jpg (deflated 3%)
  adding: kaggle/working/test/image/093.jpg (deflated 6%)
  adding: kaggle/working/test/image/060.jpg (deflated 3%)
  adding: kaggle/working/test/image/033.jpg (deflated 3%)
  adding: kaggle/working/test/image/107.jpg (deflated 7%)
  adding: kaggle/working/test/image/038.jpg (deflated 5%)
  adding: kaggle/working/test/image/043.jpg (deflated 3%)
  adding: kaggle/working/test/image/030.jpg (deflated 4%)
  adding: kaggle/working/test/image/097.jpg (deflated 6%)
  adding: kaggle/working/test/image/059.jpg (deflated 4%)
  adding: kaggle/working/test/image/050.jpg (deflated 7%)
  adding: kaggle/working/test/image/106.jpg (deflated 9%)
  adding: kaggle/working/test/image/082.jpg (deflated 3%)
  adding: kaggle/working/test/image/054.jpg (deflated 5%)
  adding: kaggle/working/test/image/017.jpg (deflated 8%)
  adding: kaggle/working/test/image/090.jpg (deflated 4%)
  adding: kaggle/working/test/image/035.jpg (deflated 9%)
  adding: kaggle/working/test/image/036.jpg (deflated 4%)
  adding: kaggle/working/test/image/062.jpg (deflated 3%)
  adding: kaggle/working/test/image/064.jpg (deflated 4%)
  adding: kaggle/working/coco_eval.py (deflated 76%)
  adding: kaggle/working/utils.py (deflated 70%)
  adding: kaggle/working/.virtual_documents/ (stored 0%)
  adding: kaggle/working/val.json (deflated 80%)
  adding: kaggle/working/submission.csv (deflated 62%)

	ID	category	Width	Height	Filename	ImagePath	MarkPath
0	TDR04_20180315_075734	normal	2328	2344	220_97.dcm	/normal/image/220_97.dcm	/normal/mark/220_97.dcm.jpg
1	TDR04_20180315_080518	normal	2472	2560	220_94.dcm	/normal/image/220_94.dcm	/normal/mark/220_94.dcm.jpg
2	TDR04_20180315_081322	normal	2312	2496	220_93.dcm	/normal/image/220_93.dcm	/normal/mark/220_93.dcm.jpg
3	TDR04_20180315_081746	normal	2448	2584	220_92.dcm	/normal/image/220_92.dcm	/normal/mark/220_92.dcm.jpg
4	TDR04_20180315_082113	normal	2144	2384	220_91.dcm	/normal/image/220_91.dcm	/normal/mark/220_91.dcm.jpg
...	...	...	...	...	...	...	...
446	TDR02_20161209_161439	脊椎側彎	2376	2592	4440_5.dcm	/脊椎側彎/image/4440_5.dcm	/脊椎側彎/mark/4440_5.dcm.jpg
447	TDR04_20180224_084933	脊椎側彎	2248	2600	4440_0.dcm	/脊椎側彎/image/4440_0.dcm	/脊椎側彎/mark/4440_0.dcm.jpg
448	TDR04_20180226_082354	脊椎側彎	2488	2456	4440.dcm	/脊椎側彎/image/4440.dcm	/脊椎側彎/mark/4440.dcm.jpg
449	TDR01_20171106_095308	脊椎側彎	2320	2376	A0_29.dcm	/脊椎側彎/image/A0_29.dcm	/脊椎側彎/mark/A0_29.dcm.jpg
450	TDR01_20171108_100516	脊椎側彎	2280	2288	A0_28.dcm	/脊椎側彎/image/A0_28.dcm	/脊椎側彎/mark/A0_28.dcm.jpg

	ID	category	Width	Height	Filename	ImagePath	MarkPath
0	TDR04_20180315_075734	normal	2328	2344	220_97.dcm	normal/image/220_97.dcm	normal/mark/220_97.dcm.jpg
1	TDR04_20180315_080518	normal	2472	2560	220_94.dcm	normal/image/220_94.dcm	normal/mark/220_94.dcm.jpg
2	TDR04_20180315_081322	normal	2312	2496	220_93.dcm	normal/image/220_93.dcm	normal/mark/220_93.dcm.jpg
3	TDR04_20180315_081746	normal	2448	2584	220_92.dcm	normal/image/220_92.dcm	normal/mark/220_92.dcm.jpg
4	TDR04_20180315_082113	normal	2144	2384	220_91.dcm	normal/image/220_91.dcm	normal/mark/220_91.dcm.jpg
...	...	...	...	...	...	...	...
446	TDR02_20161209_161439	scoliosis	2376	2592	4440_5.dcm	scoliosis/image/4440_5.dcm	scoliosis/mark/4440_5.dcm.jpg
447	TDR04_20180224_084933	scoliosis	2248	2600	4440_0.dcm	scoliosis/image/4440_0.dcm	scoliosis/mark/4440_0.dcm.jpg
448	TDR04_20180226_082354	scoliosis	2488	2456	4440.dcm	scoliosis/image/4440.dcm	scoliosis/mark/4440.dcm.jpg
449	TDR01_20171106_095308	scoliosis	2320	2376	A0_29.dcm	scoliosis/image/A0_29.dcm	scoliosis/mark/A0_29.dcm.jpg
450	TDR01_20171108_100516	scoliosis	2280	2288	A0_28.dcm	scoliosis/image/A0_28.dcm	scoliosis/mark/A0_28.dcm.jpg

	ID	category	Width	Height	Filename	ImagePath	MarkPath
0	TDR04_20180315_075734	normal	2328	2344	220_97.dcm	normal/image/220_97.dcm	normal/mark/220_97.dcm.jpg
80	TDR04_20180227_083423	aortic_curvature	2504	2536	220_14.dcm	aortic_curvature/image/220_14.dcm	aortic_curvature/mark/220_14.dcm.jpg
132	TDR01_20190313_090724	aortic_atherosclerosis_calcification	2392	2600	10_1d.dcm	aortic_atherosclerosis_calcification/image/10_...	aortic_atherosclerosis_calcification/mark/10_1...
203	TDR04_20180226_090403	cardiac_hypertrophy	2008	2280	4440.dcm	cardiac_hypertrophy/image/4440.dcm	cardiac_hypertrophy/mark/4440.dcm.jpg
236	TDR05_20151105_094209	intercostal_pleural_thickening	2296	2512	4440_4.dcm	intercostal_pleural_thickening/image/4440_4.dcm	intercostal_pleural_thickening/mark/4440_4.dcm...
265	TDR04_20180227_083423	lung_field_infiltration	2504	2536	220_3.dcm	lung_field_infiltration/image/220_3.dcm	lung_field_infiltration/mark/220_3.dcm.jpg
333	TDR04_20180227_085056	degenerative_joint_disease_of_the_thoracic_spine	2336	2360	220_15.dcm	degenerative_joint_disease_of_the_thoracic_spi...	degenerative_joint_disease_of_the_thoracic_spi...
393	TDR01_20171109_083459	scoliosis	2232	2408	A0_26.dcm	scoliosis/image/A0_26.dcm	scoliosis/mark/A0_26.dcm.jpg

	ID	category	Width	Height	Filename	ImagePath	MarkPath	xmin	ymin	xmax	ymax
0	TDR04_20180315_075734	normal	2328	2344	220_97.dcm	normal/image/220_97.dcm	normal/mark/220_97.dcm.jpg	0	0	0	0
1	TDR04_20180315_080518	normal	2472	2560	220_94.dcm	normal/image/220_94.dcm	normal/mark/220_94.dcm.jpg	0	0	0	0
2	TDR04_20180315_081322	normal	2312	2496	220_93.dcm	normal/image/220_93.dcm	normal/mark/220_93.dcm.jpg	0	0	0	0
3	TDR04_20180315_081746	normal	2448	2584	220_92.dcm	normal/image/220_92.dcm	normal/mark/220_92.dcm.jpg	0	0	0	0
4	TDR04_20180315_082113	normal	2144	2384	220_91.dcm	normal/image/220_91.dcm	normal/mark/220_91.dcm.jpg	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...
446	TDR02_20161209_161439	scoliosis	2376	2592	4440_5.dcm	scoliosis/image/4440_5.dcm	scoliosis/mark/4440_5.dcm.jpg	1016	560	1432	2328
447	TDR04_20180224_084933	scoliosis	2248	2600	4440_0.dcm	scoliosis/image/4440_0.dcm	scoliosis/mark/4440_0.dcm.jpg	912	552	1408	2272
448	TDR04_20180226_082354	scoliosis	2488	2456	4440.dcm	scoliosis/image/4440.dcm	scoliosis/mark/4440.dcm.jpg	1016	464	1560	2184
449	TDR01_20171106_095308	scoliosis	2320	2376	A0_29.dcm	scoliosis/image/A0_29.dcm	scoliosis/mark/A0_29.dcm.jpg	1032	512	1384	2016
450	TDR01_20171108_100516	scoliosis	2280	2288	A0_28.dcm	scoliosis/image/A0_28.dcm	scoliosis/mark/A0_28.dcm.jpg	912	592	1392	2120

	ID	category	Width	Height	Filename	ImagePath	MarkPath	xmin	ymin	xmax	ymax	class_id
0	TDR04_20180315_075734	normal	2328	2344	220_97.dcm	normal/image/220_97.dcm	normal/mark/220_97.dcm.jpg	0	0	0	0	0
1	TDR04_20180315_080518	normal	2472	2560	220_94.dcm	normal/image/220_94.dcm	normal/mark/220_94.dcm.jpg	0	0	0	0	0
2	TDR04_20180315_081322	normal	2312	2496	220_93.dcm	normal/image/220_93.dcm	normal/mark/220_93.dcm.jpg	0	0	0	0	0
3	TDR04_20180315_081746	normal	2448	2584	220_92.dcm	normal/image/220_92.dcm	normal/mark/220_92.dcm.jpg	0	0	0	0	0
4	TDR04_20180315_082113	normal	2144	2384	220_91.dcm	normal/image/220_91.dcm	normal/mark/220_91.dcm.jpg	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...
446	TDR02_20161209_161439	scoliosis	2376	2592	4440_5.dcm	scoliosis/image/4440_5.dcm	scoliosis/mark/4440_5.dcm.jpg	1016	560	1432	2328	7
447	TDR04_20180224_084933	scoliosis	2248	2600	4440_0.dcm	scoliosis/image/4440_0.dcm	scoliosis/mark/4440_0.dcm.jpg	912	552	1408	2272	7
448	TDR04_20180226_082354	scoliosis	2488	2456	4440.dcm	scoliosis/image/4440.dcm	scoliosis/mark/4440.dcm.jpg	1016	464	1560	2184	7
449	TDR01_20171106_095308	scoliosis	2320	2376	A0_29.dcm	scoliosis/image/A0_29.dcm	scoliosis/mark/A0_29.dcm.jpg	1032	512	1384	2016	7
450	TDR01_20171108_100516	scoliosis	2280	2288	A0_28.dcm	scoliosis/image/A0_28.dcm	scoliosis/mark/A0_28.dcm.jpg	912	592	1392	2120	7

這份檔案會講解以下幾個部分:¶

Rename files & Save .csv as UTF-8 encoding format¶

Check data & images¶

更改類別名稱為英文¶

畫出8種類別的第一張 image & mask 位置¶

X-ray image normalization¶

Mask image to bounding box¶

Write class id¶

Split training set and validation set¶

Dataset to COCO format¶

Format:¶

Images¶

Annotations¶

Categories¶

Bounding boxes format¶

Save files¶

TO DO:¶

hw5-3¶

5-2¶